What was tested
Rajasekar and colleagues in Cynthia Yiu’s group at the University of Hong Kong Faculty of Dentistry, with co-authors at the University of Alberta, Al-Azhar University and Umm Al-Qura University, published in Stem Cell Research & Therapy (volume 16, article 34, 4 February 2025; open access, CC BY-NC-ND 4.0; funded by an HKU Seed Fund for Basic Research grant, project 202111159148). The question: do peptide-based glycogen synthase kinase-3β (GSK-3β) inhibitors, used as pseudosubstrates, drive odontogenic differentiation in human dental pulp stem cells (hDPSCs), and how do they compare with tideglusib, the best-known small-molecule GSK-3β inhibitor in dentin-repair research? The four peptides were bought from GL Biochem in Shanghai and were all previously published tools, not new chemistry: pS9 (mimics the N-terminal autoinhibitory Ser9 region of GSK-3β itself), LRP6a (mimics the GSK-3β binding site on the LRP6 receptor), L803 (derived from the substrate heat shock factor 1), and L803-mts, which is L803 with a myristoyl group added to make it cell-permeable. The cells came from the caries-free maxillary third molars of five donors aged 25 to 30 (HKU IRB approval 20221026-003-000), used at passages 3 to 5. Cells in odontogenic medium alone were the negative control; odontogenic medium plus tideglusib at 100 nM was the positive control. Readouts were viability and proliferation, alkaline phosphatase (ALP) at day 14, alizarin red mineralization at day 21, qPCR and western blotting for odontogenic markers, β-catenin immunofluorescence and cytoplasmic/nuclear fractionation westerns, and a transwell migration assay.
What they found
All four peptides were non-toxic to hDPSCs at concentrations up to 200 µM and did not change proliferation. On ALP at day 14, all four peptides at 100 µM beat the negative control, but only L803-mts significantly beat the tideglusib positive control: 14.9 nmol pNPP per µg of total protein versus 10.6 for tideglusib (p < 0.05). The day 21 alizarin red readout matched that pattern: 100 µM L803-mts significantly out-mineralized both controls. By qPCR, ALP mRNA was significantly up at day 14, and by day 21 DSPP, DMP-1, Runx-2 and OPN were significantly increased, with OCN measured but not reported as significant. Western blots at day 21 showed a significant DMP-1 protein increase, while the DSPP protein increase was moderate and not significant (p > 0.05), a detail that matters below. β-catenin accumulated in the nucleus by both immunofluorescence and fractionation, and 100 µM L803-mts significantly increased cell migration across the transwell. The authors state this is the first study to comprehensively evaluate peptide-based GSK-3β inhibitors on hDPSCs.
What it does not show
Everything is in dishes: no animal work, no reparative dentin formed in a tooth, no inflammatory environment resembling pulpitis. No enzyme assay was run, so the paper reports no IC50 or Ki for any peptide; potency rests entirely on cellular readouts. The tideglusib comparison used a single fixed concentration of 100 nM with no dose-matching, so “beats tideglusib” is bounded to these assay conditions and says nothing about which agent would perform better at an injury site. The Wnt mechanism is asserted rather than proven: there is no TOPFlash reporter, no siRNA knockdown and no rescue experiment with a Wnt pathway inhibitor, so nuclear β-catenin accumulation is consistent with, but does not establish, canonical Wnt signaling as the driver of differentiation. The abstract claims increased protein expression of DSPP and DMP-1, but the results text reports the DSPP protein increase as not significant; we follow the results text. One framing in the paper, that small-molecule GSK-3β inhibitors act irreversibly, is inaccurate (tideglusib, BIO and CHIR99021 are reversible ATP competitors) and we do not repeat it. A production anomaly: the competing-interests section of the article contains only the word “Funding:”, so there is no usable conflict-of-interest statement, and we report none rather than assuming one. As the authors themselves note, cells from five young healthy donors may not reproduce in vivo behavior, and the peptides’ susceptibility to proteases at an injury site is an open delivery problem.
Where it sits in the field
For the pulp and dentine repair program, this is a tool comparison inside the GSK-3β lane, not a tier input. The interest for the record is the head-to-head design: most dentin-repair papers test one agent against a blank control, so a peer-reviewed, if assay-level, data point showing a peptide inhibitor exceeding tideglusib on mineralization under single-concentration conditions is a useful reference when the small-molecule route’s ceiling is debated. The caveats cut the other way too: these peptides are roughly two-decade-old research tools, the dose translation from 100 µM in culture to a capping material is unexplored, and the missing Wnt rescue leaves the mechanism where this field has been burned before by correlation-as-causation. No program tier moves on an in vitro result; the current field assessment stands at /field/.
Where we differ from the coverage
We found no press or popular coverage of this paper to differ from. Against the paper’s own abstract, we note the DSPP protein discrepancy above, and we decline to repeat its irreversible-inhibition framing about small molecules. The “outperforms tideglusib” claim is real at the level of these assays under single-concentration conditions; it is not evidence of superior pulp capping in an animal or a person, and nothing in the paper tests delivery, stability or safety in a tooth.
Provenance: grounded in the full open-access text (CC BY-NC-ND 4.0) of Rajasekar et al., Stem Cell Research & Therapy 16:34, 2025, DOI 10.1186/s13287-025-04150-7, read in full from the Europe PMC version PMC11792195; every number above was checked against that text. Exact qPCR fold-changes and mineralization percentages appear only in figures and are not quoted here. Method and sourcing standard at /method/.